树莓派 2 到手 性能提升明显

周五拿到的，发现公司没有hdmi显示器，昨天把镜像写到u盘发现不能从u盘启动，jd撸了张三棒子的U1的TF，写完镜像发现官方ubuntu默认ssh是关闭的，换了树莓派大便好歹连上了。

正好用上次在张大妈180剁到号称SSD的U盘来做测试 SanDisk Extreme CZ80 NTFS

写入测试
root@raspberrypi:/mnt/sandisk#
root@raspberrypi:/mnt/sandisk# dd bs=64k count=1k if=/dev/zero of=test oflag=dsync
1024+0 records in
1024+0 records out
67108864 bytes (67 MB) copied, 4.62741 s, 14.5 MB/s
root@raspberrypi:/mnt/sandisk# dd bs=128k count=1k if=/dev/zero of=test oflag=dsync
1024+0 records in
1024+0 records out
134217728 bytes (134 MB) copied, 9.35676 s, 14.3 MB/s
root@raspberrypi:/mnt/sandisk# dd bs=1M count=1K if=/dev/zero of=test oflag=dsync
1024+0 records in
1024+0 records out
1073741824 bytes (1.1 GB) copied, 77.8103 s, 13.8 MB/s
root@raspberrypi:/mnt/sandisk#

峰值负载
top - 15:57:39 up 19 min, 3 users, load average: 1.05, 0.45, 0.24
Tasks: 91 total, 2 running, 89 sleeping, 0 stopped, 0 zombie
%Cpu(s): 11.5 us, 16.0 sy, 0.0 ni, 59.3 id, 12.9 wa, 0.0 hi, 0.3 si, 0.0 st
KiB Mem: 949328 total, 932412 used, 16916 free, 438872 buffers
KiB Swap: 102396 total, 0 used, 102396 free, 447876 cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
2292 root 20 0 4592 2468 1928 R 87.7 0.3 1:17.07 mount.ntfs
2298 root 20 0 4608 2524 1472 S 13.9 0.3 0:10.34 dd



复制1073741824 bytes (1.1 GB) NTFS to NTFS用时2分半

Sat Mar 7 16:05:53 UTC 2015
Sat Mar 7 16:08:32 UTC 2015

峰值负载
top - 16:07:37 up 29 min, 3 users, load average: 1.37, 0.77, 0.45
Tasks: 94 total, 2 running, 92 sleeping, 0 stopped, 0 zombie
%Cpu(s): 7.3 us, 12.5 sy, 0.0 ni, 73.1 id, 7.0 wa, 0.0 hi, 0.1 si, 0.0 st
KiB Mem: 949328 total, 936724 used, 12604 free, 441784 buffers
KiB Swap: 102396 total, 0 used, 102396 free, 450232 cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
2292 root 20 0 4592 2468 1928 R 63.2 0.3 3:41.17 mount.ntfs
2327 root 20 0 3924 1228 1096 S 10.9 0.1 0:09.03 cp



内网FTP测试

百兆的路由器，笔记本只有无线网卡，NTFS读取还是挺快的，如果是千兆环境应该还能再快一点，应该能到16M/s附近。

[右] 227 Entering Passive Mode (192,168,61,117,84,193).
[右] 正在打开数据连接 IP: 192.168.61.117 端口: 21697
[右] RETR 1.mp4
[右] 150 Opening BINARY mode data connection for 1.mp4 (132995681 bytes).
[右] 226 Transfer complete. 下载: 1.mp4 126.83 MB 于 14 秒 (9.29 MB/秒)

系统负载
top - 18:53:41 up 7 min, 1 user, load average: 0.05, 0.13, 0.08
Tasks: 83 total, 1 running, 82 sleeping, 0 stopped, 0 zombie
%Cpu(s): 0.0 us, 1.8 sy, 0.0 ni, 97.3 id, 0.0 wa, 0.0 hi, 0.9 si, 0.0 st
KiB Mem: 949328 total, 928468 used, 20860 free, 446180 buffers
KiB Swap: 102396 total, 0 used, 102396 free, 451412 cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
2205 root 20 0 4468 1748 1436 S 4.3 0.2 0:11.24 vsftpd
3 root 20 0 0 0 0 S 1.3 0.0 0:01.83 ksoftirqd/0


Benchmarks
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)

System: raspberrypi: GNU/Linux
OS: GNU/Linux -- 3.18.7-v7+ -- #755 SMP PREEMPT Thu Feb 12 17:20:48 GMT 2015
Machine: armv7l (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: ARMv7 Processor rev 5 (v7l) (0.0 bogomips)

CPU 1: ARMv7 Processor rev 5 (v7l) (0.0 bogomips)

CPU 2: ARMv7 Processor rev 5 (v7l) (0.0 bogomips)

CPU 3: ARMv7 Processor rev 5 (v7l) (0.0 bogomips)

16:59:18 up 5 min, 1 user, load average: 0.08, 0.06, 0.03; runlevel 2

------------------------------------------------------------------------
Benchmark Run: Sat Mar 07 2015 16:59:18 - 17:27:29
4 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 2956203.9 lps (10.0 s, 7 samples)
Double-Precision Whetstone 498.4 MWIPS (9.9 s, 7 samples)
Execl Throughput 363.9 lps (29.6 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 73543.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 21005.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 193476.5 KBps (30.0 s, 2 samples)
Pipe Throughput 175674.7 lps (10.0 s, 7 samples)
Pipe-based Context Switching 31675.2 lps (10.0 s, 7 samples)
Process Creation 1287.5 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 1193.2 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 335.5 lpm (60.1 s, 2 samples)
System Call Overhead 410617.8 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 2956203.9 253.3
Double-Precision Whetstone 55.0 498.4 90.6
Execl Throughput 43.0 363.9 84.6
File Copy 1024 bufsize 2000 maxblocks 3960.0 73543.0 185.7
File Copy 256 bufsize 500 maxblocks 1655.0 21005.5 126.9
File Copy 4096 bufsize 8000 maxblocks 5800.0 193476.5 333.6
Pipe Throughput 12440.0 175674.7 141.2
Pipe-based Context Switching 4000.0 31675.2 79.2
Process Creation 126.0 1287.5 102.2
Shell Scripts (1 concurrent) 42.4 1193.2 281.4
Shell Scripts (8 concurrent) 6.0 335.5 559.1
System Call Overhead 15000.0 410617.8 273.7
========
System Benchmarks Index Score 173.7

------------------------------------------------------------------------
Benchmark Run: Sat Mar 07 2015 17:27:29 - 17:55:45
4 CPUs in system; running 4 parallel copies of tests

Dhrystone 2 using register variables 11822874.0 lps (10.0 s, 7 samples)
Double-Precision Whetstone 1989.6 MWIPS (10.0 s, 7 samples)
Execl Throughput 1364.0 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 116583.7 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 33022.0 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 316438.0 KBps (30.0 s, 2 samples)
Pipe Throughput 700798.4 lps (10.0 s, 7 samples)
Pipe-based Context Switching 123270.4 lps (10.0 s, 7 samples)
Process Creation 2867.2 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 2690.9 lpm (60.1 s, 2 samples)
Shell Scripts (8 concurrent) 353.7 lpm (60.4 s, 2 samples)
System Call Overhead 1584507.7 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 11822874.0 1013.1
Double-Precision Whetstone 55.0 1989.6 361.7
Execl Throughput 43.0 1364.0 317.2
File Copy 1024 bufsize 2000 maxblocks 3960.0 116583.7 294.4
File Copy 256 bufsize 500 maxblocks 1655.0 33022.0 199.5
File Copy 4096 bufsize 8000 maxblocks 5800.0 316438.0 545.6
Pipe Throughput 12440.0 700798.4 563.3
Pipe-based Context Switching 4000.0 123270.4 308.2
Process Creation 126.0 2867.2 227.6
Shell Scripts (1 concurrent) 42.4 2690.9 634.7
Shell Scripts (8 concurrent) 6.0 353.7 589.4
System Call Overhead 15000.0 1584507.7 1056.3
========
System Benchmarks Index Score 443.7

======= Script description and score comparison completed! =======

总体来看秒一代十几条街不成问题。虽然网卡/USB仍然是共享480M带宽，NTFS的情况来看主要是写入有瓶颈，主要在CPU上（严格的来说在ntfs驱动上）用来做家庭NAS还是比较合适的，应该能跑满百兆带宽。做离线下载或者PT应该很生猛的，挂个几十TB种子问题应该也不大。如果拿来做web之类的开发问题应该也不大，跑个环境GUI+Editor应该绰绰有余的。